Image fusion for interventional guidance

The method of claim 3, further comprising detecting said global position, anatomical landmarks, and first and second contours are using marginal space learning with a hierarchical approach, wherein detectors are successively trained using probabilistic boosting trees.

A method of transforming target structure anatomies in a pre-operative image I2 into an intra-operative image I1, comprising the steps of:

• determining a transformation Φ aligns a target structure T2 and an anchor structure A2 in the pre-operative image I2 into a corresponding target structure T1 and anchor structure A1 in the intra-operative image I1 by finding a transformation {circumflex over (Φ)} that maximizes a functional log(P(Φ|I1, A2)) using an expectation-maximization approach, wherein the target structure T1 is not visible in the intra-operative image.

A non-transitory program storage device readable by a computer, tangibly embodying a program of instructions executed by the computer to perform the method steps for transforming target structure anatomies in a pre-operative image I2 into an intra-operative image I1, the method comprising the steps of:

• determining a transformation Φ aligns a target structure T2 and an anchor structure A2 in the pre-operative image I2 into a corresponding target structure T1 and anchor structure A1 in the intra-operative image I1 by finding a transformation {circumflex over (Φ)} that maximizes a functional log(P(Φ|I1, A2)) using an expectation-maximization approach, wherein the target structure T1 is not visible in the intra-operative image.